基于FOA-GRNN的纳米铁粉分解炉温度预测

发布时间：2018-06-21 09:09

  本文选题：纳米铁粉 + 温度预测　； 参考：《中国测试》2017年04期

【摘要】：为提高纳米铁粉的制备工艺,实现纳米铁粉分解炉温度的精确控制,提出一种基于果蝇优化算法和广义回归神经网络的纳米铁粉分解炉温度预测方法。该方法采用现场采集数据,选取进液量和各个温区加热装置的开度因素来预测待预测温区温度。通过广义回归神经网络,建立温度预测模型,并利用果蝇优化算法对光滑因子进行动态寻优。选取不同种群规模对建立模型进行验证,并将该文建立模型与普通广义神经网络和粒子群算法优化的广义神经网络模型的预测效果进行对比。验证表明:该文建立模型平均相对误差为0.43%,且能够排除人为设置参数的干扰,具有较好的准确性与稳定性,可进一步用于分解炉温度控制的研究。
[Abstract]:In order to improve the preparation process of nanometer iron powder and realize the accurate control of the temperature of nanometer iron powder decomposing furnace, a temperature prediction method of nanometer iron powder decomposing furnace based on Drosophila optimization algorithm and generalized regression neural network was proposed. In this method, the field data are collected, and the temperature of the temperature region is predicted by selecting the input liquid quantity and the opening factor of the heating device in each temperature zone. The temperature prediction model was established by generalized regression neural network, and the smoothing factor was dynamically optimized by Drosophila optimization algorithm. Different population sizes are selected to verify the model, and the prediction results of this model are compared with that of the generalized neural network model and particle swarm optimization model based on general generalized neural network (GNN) and particle swarm optimization (PSO). The results show that the average relative error of the model is 0.43 and the disturbance of artificial parameters can be eliminated. The model has good accuracy and stability and can be further used in the study of calciner temperature control.
【作者单位】： 长春工业大学电气与电子工程学院;
【基金】：吉林省重点科技攻关项目(20140204024GX)
【分类号】：TP18;TB383.1
，

本文编号：2048011